What’s the weirdest technical request you’ve ever received from your product team? We’ve had a few, but the most extreme by far was this: generate an entire book from existing data—without human intervention.
This wasn’t just a fun side project; it was a real production need, requiring precision, automation, and some unexpected technical creativity.
In this talk, we’ll explore:
The unique challenges of programmatic publishing and why writing a book with code is harder than it sounds.

How we built a millimeter perfect system that operates at scale and the libraries that made it possible: Docker, AWS Lambda, jsPdf, serverless.js, and one very surprising tool :).

How we leveraged AI to assist in the process—so we wrote an entire component in a language we don’t know.

Using js proxy to add support to fluent syntax
How we fell for a bug that once crashed a space-vehicle, and learned to revalidate our assumptions.

How we carefully selected the right library for each part of the book and turned the project into a production-ready.

This talk is a deep dive into the intersection of Node.js, automation, and publishing, packed with insights that can be applied to any large-scale content generation problem. We’ll share what worked, what didn’t, and how Node.js enabled us to implement a unique project and led us to solutions we never expected.